Assessments of the decadal predictability of Atlantic-European climate and development of the European capability for decadal climate forecasting
The predictability of decadal variability has been investigated by carrying out "perfect ensemble" experiments in which the oceanic initial conditions are assumed perfectly known, but the atmospheric initial conditions are perturbed. The results indicate that in general the strength of the Atlantic Thermohaline Circulation (THC) is potentially predictable at least a decade in advance and, in some situations, multi-decadal predictions of the THC may be possible. In addition, THC-related variations in sea surface temperature and surface air temperature are potentially predictable one or two decades in advance. The exact level of predictability is dependent on the oceanic initial conditions and on the coupled model used.
The influence of the ocean on the climate of the North Atlantic European region has been extensively investigated through the use of coordinated experiments with four different atmospheric General Circulation Models. Multidecadal ensemble simulations were carried out in which the models were forced with a reconstruction of historical variations in sea surface temperatures (SST). The ensembles were first used to quantify the potential predictability of climate fluctuations. This analysis revealed that potential decadal predictability is highest in the summer season both for tropical and extra-tropical parts of the North Atlantic European (NAE) region. In summer (winter), roughly 60% (50%) and 30% (20%) of the variance is potentially predictable for the tropical and extra-tropical parts of the NAE region respectively. There are, however, significant differences between estimates of potential predictability from different atmosphere models, particularly in spring and autumn. An optimal detection methodology was applied to the ensemble simulations to determine the space-time characteristics of the oceanic influence on NAE climate. It has been shown that the ocean exerts an important influence on multidecadal timescales as well as on interannual timescales. Multidecadal variations in Atlantic SST, which may be driven by the thermohaline circulation, modulate European climate. The pattern of the atmospheric response in winter has a strong projection on the North Atlantic Oscillation pattern. On interannual timescales NAE climate is influenced by ENSO but also by Atlantic SST. Notably, PREDICATE results indicate that the relative importance of these two influences is modulated by the multidecadal variation of Atlantic SST. To clarify the specific role of the Atlantic ocean in NAE climate additional experiments were carried out in which the models were forced by idealised patterns of Atlantic SST anomalies. The SST anomalies were identified from observations as those most likely to induce a significant response. A key finding is that, contrary to expectations, the response to the SST forcing is very consistent between the different atmosphere models. In many cases, the uncertainty is significantly less than the signal strength. The magnitude of the response is generally smaller than the interannual variability but is sufficient to be of clear importance for understanding and predicting decadal variability.
Simulations of the evolution of the Atlantic Ocean during the period 1958-98 were carried out using five different ocean general circulation models forced with daily surface fields derived from the NCEP reanalysis. For two of the models ensemble simulations were carried out to investigate the role of the oceanic initial conditions. Most of the interannual-to-decadal variability of the Atlantic Meridional Overturning Circulation (MOC) is determined by the common surface forcing. The results suggest that the strength of the MOC and associated heat transport have varied considerably over the last 50 years, with a decreasing trend of 1-2Sv and 0.15PW between 1950-1960, and a 3.5-4.5Sv and 0.2PW increase since 1960. This variability is caused by density anomalies in the North Atlantic subpolar gyre. The North Atlantic Oscillation plays a key role in the formation of these anomalies by modulating air-sea fluxes in the northwest Atlantic, leading to anomalous mixing. Changes in the MOC lag changes in mixing by a few years. The ensemble simulations indicate that the evolution of the MOC can be sensitive to changes in the oceanic initial conditions. This influence is mainly seen on multidecadal timescales and appears to be linked to the salinity field in the high latitude North Atlantic. Further simulations have been used to explore sensitivity to resolution. The results show that the representation of the formation, propagation and decay of key observed, large-scale, dynamic and thermodynamic anomalies in the North Atlantic Ocean depends on model resolution, and that the major anomalies can be realistically simulated with model resolutions of 40km or less.